A GLOBAL TEST OF A TIME-FREQUENCY SMALL-EVENT DISCRIMINANT

The Comprehensive nuclear-Test-Ban Treaty (CTBT), which was recently a dopted by the UN General Assembly and signed by President Clinton, has lowered the testing yield limit to zero and raised the profile of sma ll earthquakes and mining blasts. Because relatively small events are very common, there is a strong need for automated algorithms that can be used to screen out the events that are obviously chemical or natura l and identify those few curious enough to warrant closer scrutiny. Th e primary objective of this article is to assess the utility of high-f requency spectral modulations for the discrimination of mining blasts from earthquakes and single explosions at near-regional distances. Min ing blasts commonly yield spectral modulations that are independent of time and the recording component. This article describes an automated discriminant that looks for these delay-fire diagnostics in data reco rded on one or three components by single stations, arrays, or regiona l networks. Distinct deployments in central Asia, Europe, and North Am erica are used to assess the transportability of the approach. The dis criminant tests give misclassification probabilities, estimated with m ultivariate statistics, ranging from 0.5 to 3.5%. Discrimination using time-frequency expansions does not rely on expert interpretation but is quite routine. The article explores likely causes of the occasional discrimination outliers. Factors that can eliminate spectral modulati ons from a delay-fired event include attenuation, detonation anomalies (where deviations from the designed, regular, shot sequence occur), a nd waveform variability. Some natural events and single explosions wil l exhibit spectral modulations that most likely result from propagatio n resonance. Data from Kyrgyzstan and Nevada is used to illustrate the se effects; however, inadequate ground truth information and lack of c alibration explosions inmost of these datasets keeps definitive conclu sions, regarding why the method will sometimes fail, out of reach. The se observations underscore the need to train this algorithm to most ef fectively deal with these processes and pair it with other, complement ary, discriminants to allow accurate characterization of all small eve nts.